Make a Hotel Lobby AI video from two photos
Prepare two portraits, cast the orange-booth roles, and review a Hotel Lobby video remake before adding your final soundtrack.
The Hotel Lobby setup works as a short visual duet: two people share an orange booth and a hanging microphone. In CastRemix, you supply their photos and the template supplies the performance reference. Start with the Hotel Lobby template and decide who belongs on each side before uploading.
Prepare two separate portraits
Choose one person per image. A similar head angle and roughly comparable face size make the pair easier to judge together. A waist-up portrait can give useful information about shoulders and clothing without shrinking the face into the background. Avoid a group photo that leaves the intended person unclear.
The photos do not need to have been taken together. They do need to make each person recognizable. If one is a distant holiday snapshot and the other a close selfie, prepare more compatible crops before starting. Keep enough shoulder space for the medium two-shot.
Cast the booth, then check the previews
Photo 1 supplies the left performer; Photo 2 supplies the right performer. Check both uploaded previews against that plan. For a birthday duet, for example, put the birthday person in the chosen lead position deliberately rather than trusting the order of filenames.
The motion video is already included. This fixed template also supplies its instructions, duration and landscape framing. Choose an available resolution and inspect the current quote. You do not need to paste a prompt or upload a second copy of the booth clip.
Inspect the microphone and the shared space
Review both people throughout the result. Pay attention to the moment a head turns near the shared microphone: a face can remain plausible while merging with the mic or the other person. Check that clothing stays attached to the same person and that the two bodies remain distinct.
Compare the opening, middle and ending, then watch at normal speed. A steady face in one still does not establish that every gesture works. Reference-led generation can change the staging, so treat the preview as an input to follow rather than proof of an exact final match.
Add sound as a separate creative decision
This template is a visual performance remake. It does not write a personalized rap, record a new vocal or promise exact lip sync. If your idea depends on a specific line, plan that audio separately and assess whether the generated mouth movement fits the use.
For a greeting, a short spoken introduction outside the performance can work better than forcing new lyrics onto an existing mouth pattern. Use audio you can publish and check the final mix after downloading. The music workflow explains how to keep the original track separate from the visual draft.
Make the next attempt answer one question
If the wrong people appear on the wrong sides, correct the slots first. If only one face drifts, change that person's photo while keeping the other reference fixed. Repeating the same setup without identifying a problem gives you little to compare.
For alternatives to the booth, explore duo video ideas. For a closer look at casting, use the two-photo role guide.